The improvement and knowledge-transfer division of the AGL Failure Database. Divisions 01–20 establish what failed, Division 21 establishes why, Division 22 establishes how serious it was, and Division 23 establishes the reliability trend it creates. Division 24 closes the loop: what was done about it, how recurrence is prevented, and what was learned — the closed-loop improvement system that turns individual failures into organizational engineering knowledge.
Key principle — Do not close a failure merely because the equipment has been repaired. Close it only when the required corrective action has been verified, recurrence risk has been addressed, and the knowledge gained has been captured for future maintenance, design, procurement, testing and projects. Corrective guides, preventive guides and lessons-learned reports become controlled outputs of this division rather than standalone documents.
The action taken immediately after failure detection to control the situation, restore required functionality or reduce immediate risk.
Immediate corrective action does not necessarily eliminate the root cause.
A short-term action used to restore or maintain functionality until a permanent solution can be implemented.
Temporary restoration must not automatically be treated as failure closure.
The action that addresses the established failure mechanism/root cause and restores the equipment/system to the required condition.
Division 21 Root Cause → Division 24 Permanent Corrective Action. The permanent action should be demonstrably related to the identified cause.
Replacement of a failed or degraded component.
Component replacement records can later feed Division 23 — Component Reliability Analysis.
Repair of failed equipment rather than complete replacement.
Restoration of the complete affected AGL function/system after failure — broader than replacing one component.
CCR failure
Testing performed after corrective action to confirm that the repaired/restored system meets the applicable requirements.
Corrective Action → Retesting → Verification. Retesting should be recorded rather than simply marking a repair as completed.
Formal closure of the failure after corrective action and required verification have been completed.
A failure should normally not be closed until:
Modification of preventive-maintenance tasks based on failure evidence.
Repeated or significant failures identified through Divisions 21–23.
Changing inspection frequency based on demonstrated failure behaviour or risk.
Recurring connector failures:
Frequency changes should be supported by:
Modification of the design to eliminate or reduce a recurring failure mechanism.
This is one of the most valuable outputs of the Failure Database.
Modification of installation practices or installation requirements to prevent recurrence.
Potentially update:
Improvement or substitution of an unreliable component based on reliability evidence.
This creates a direct link with Division 19 → Division 23 → Division 24.
Modification of training or competency requirements based on identified human or procedural failure patterns.
Repeated failures associated with:
Revision or creation of procedures to prevent recurrence.
The database should record:
Modification of testing requirements to improve failure detection or verify corrective action.
Some failures occur because the original testing regime did not detect the developing condition.
Improvement of equipment protection against environmental conditions identified as failure contributors.
Modification of spare-parts planning based on actual failure and consumption data.
Spare strategy should consider both failure frequency and criticality. A rarely failing component may still require strategic stock if its failure is safety-critical and has a long procurement lead time.
The general lesson learned from the failure.
Repeated moisture-related failures indicate that visual inspection alone may not be sufficient to identify developing sealing problems.
The lesson should be specific enough to influence future engineering or maintenance decisions.
A lesson concerning installation workmanship, installation sequence, equipment positioning or installation requirements.
A lesson learned from testing or commissioning.
A lesson related to inspection, troubleshooting, repair or preventive maintenance.
A lesson that should influence future AGL design.
Design lessons should be fed into future project design reviews and technical specifications.
A lesson concerning product manufacturing, assembly, quality or production controls.
A lesson that improves the organisation's understanding of long-term equipment reliability.
This should feed directly into Division 23.
A lesson concerning aviation safety, electrical safety, operational safety or safety-critical system behaviour.
Safety lessons should receive high visibility in the knowledge-management system.
A recommendation that should be incorporated into future AGL projects, specifications, procurement, installation, commissioning or maintenance.
This is where the Failure Database becomes a future-project engineering knowledge base.
Standard CAPA record structure, corrective-vs-preventive distinction, effectiveness verification, workflow and knowledge-loop that apply across all 27 Division 24 topics.
Every significant failure should carry a dedicated CAPA-style record with the following fields:
Identification
Corrective Action
Preventive Action
Lessons Learned
Verification
This distinction should be maintained carefully.
Fix the existing failure/problem.
Failed connector → replace connector → test circuit → restore service.
Prevent the same or similar failure from recurring.
Repeated connector failures → identify moisture/sealing problem → revise installation/sealing procedure → modify inspection → train personnel → monitor recurrence.
A repair is not automatically a preventive action.
One of the most important controls in Division 24. After corrective/preventive action, ask:
The most powerful feature of Division 24:
This is how the database evolves from a failure register into an AGL engineering knowledge system.
Feedback Loop
Do not close a failure merely because the equipment has been repaired. Close it only when the required corrective action has been verified, recurrence risk has been addressed, and the knowledge gained has been captured for future maintenance, design, procurement, testing and projects. This makes Division 24 the continuous-improvement and organizational-learning layer of the AGL Failure Database, while the source-supported deliverables — corrective guides, preventive guides and lessons-learned reports — become controlled outputs of the database rather than standalone documents.
Quick answers about corrective action, preventive action and lessons learned, as covered in Division 24 of the AGL Failure Database.
Division 24 is the closed-loop improvement system of the database. It contains 27 published topics (CAPA-01 to CAPA-27) covering corrective action, preventive action and lessons learned.
They fall into three groups: Corrective Action (8 topics), Preventive Action (10) and Lessons Learned (9).
Immediate, temporary and permanent corrective action, component replacement, equipment repair, system restoration, retesting and failure closure.
Preventive actions include preventive-maintenance and inspection-frequency modification, design modification, installation, component, training, SOP and testing improvement, environmental protection improvement and spare-parts strategy improvement. Lessons cover failure, installation, testing, maintenance, design, manufacturing, reliability and safety lessons, plus future-project recommendations.
Corrective action fixes the existing failure, for example replacing a failed connector, testing the circuit and restoring service. Preventive action prevents the same or similar failure recurring, for example finding a moisture or sealing problem behind repeated connector failures, revising the installation procedure, modifying inspection, training personnel and monitoring recurrence. A repair is not automatically a preventive action.
After action is taken, the record asks whether it was implemented, whether the root cause was addressed, whether the equipment was tested, whether the failure recurred, whether the failure rate improved, whether the same failure occurred elsewhere and whether further action is required.
A failure should not be closed merely because the equipment has been repaired. It should be closed only when the required corrective action has been verified, recurrence risk has been addressed and the knowledge gained has been captured for future maintenance, design, procurement, testing and projects.
The failure occurs, followed by immediate corrective action, temporary restoration if needed, RCA (Division 21), severity and criticality (Division 22) and reliability trend (Division 23). Permanent corrective action, retesting, system restoration and verification follow, then preventive action, lessons learned, SOP, design, training or product updates, future-project recommendations, recurrence monitoring and failure closure.
A corrective action closes a report — but a genuinely effective one addresses the actual root cause and prevents recurrence — real competence means understanding the system well enough to verify the fix, not just document that one was applied.
Take that knowledge further with Precision AGL Academy.
Learn how to connect:
Develop structured knowledge across AGL fundamentals, components, electrical systems, testing, photometry, troubleshooting, maintenance and commissioning.
Find the failure. Understand the cause. Build the engineering skill.
Please read before relying on any content on this page or on any record downloaded from it.